BOTS, BYTES AND BIG DATA
Artificial intelligence is making its presence felt in medicine, helping to diagnose illnesses and monitor critical cases. But can machine learning catch what doctors can’t?
How many doctors does it take to tell you how you’re doing? The answer could soon be, none.
Scientists and researchers across India are exploring the applications of artificial intelligence in healthcare — from helping diagnose illnesses to monitoring critical care.
“Artificial intelligence (AI)— or cyber-physical systems, as I like to call them — can collect digitised data or generate data, analyse and make decisions based on it,” says professor Ashutosh Sharma, secretary of the union government’s department of science and technology.
“A big advantage of AI in healthcare is that it can help where there is a scarcity of human resources, which is the case in many rural areas,” adds Dr P Anandan, CEO at the Wadhwani Institute for Artificial Intelligence (Wadhwani AI). “AI feeds on examples and patterns buried in data, and this could be a huge help in areas where there is a shortage of medical expertise. Even health workers with not much experience can benefit.”
AI in healthcare will be one of the key focus areas of the Wadhwani institute, which was inaugurated by the prime minister in Mumbai last week. It is India’s first research institute focused on artificial intelligence (AI), with potential applications in the fields of education, agriculture and infrastructure too.
Intelligent technology can take services to populations that have until now been under-served, Sharma says. “For example, we do not have enough pathologists in the country and systems that are able to effectively study a large number of samples would fill this gap.”
SigTuple, a startup that focuses on AI in diagnostics, is doing just this.
A device designed by SigTuple creates a digital image of every blood slide submitted, and analyses it. “The analysis by our algorithm in clinical trials has been found to be as accurate as a pathologist studying it,” says Rohit Kumar Pandey, co-founder and CEO of SigTuple. “Given that there are just 19,000 pathologists in the country, this is a crucial time-saver.”
BABY STEPS
At Max Healthcare, Delhi, AI is being used to help monitor critical patients—freeing up beds in the ICUs and cutting down critical care costs to patients by 30%
“Sometimes a patient is stable enough to be moved out of the ICU, but still needs very close monitoring, so doctors decide to keep them in,” says clinical director Dr Sandeep Buddhiraja. “Our pilot project shows that we could move these patients to a normal ward where the new system would monitor them 24x7 as effectively, and alert nurses or doctors as needed. With every trial case, the system is learning when to kick-start which response. And we are finding that it catches changes in patients’ parameters very early, enhancing patient care.”
THE POWER OF BIG DATA
At Wadhwani AI the aim is to collaborate with governmental, non-governmental and public sector institutions to gain access to data fields from different communities and regions.
“We’ve been in touch with the Maharashtra government’s public health department to identify which communities to focus on, and the specific problems that different villages have when it comes to healthcare. It helps in data collection,” Anandan adds.
Using such data, for example, an app can be developed which can serve as an assistant to health workers by collating the medical history of patients — from age to prior health issues, nutrition levels, living conditions, family habits, and such, data can be used for predictive analysis too, Anandan says. “We can also work on forecasting by gathering and analysing data on environmental factors in different regions to foresee, for instance, a spate of dengue cases, or help new mothers improve nutritional intake by tracking regular inputs,” he adds.
To expand its scope and impact, the institution plans to collaborate with private companies and scientists from around the world. “We want to become a hub for cross-country collaborations,” he says.“The success of an institution like ours in the field of healthcare can be measured by how much technology has entered healthcare and whether it has managed to make a difference,” Anandan says.
“Did it decrease mortality rates; could it help prevent an outbreak of an epidemic? We’ll know as we continue to progress.”